Help build ReacherX

An invitation to developers, designers, and testers: learn one part of the app and help make it easier to trust and use.

SalmanFounder, ReacherX

Jul 15, 20263 min readUpdated Sep 11, 2026

Building solo gets lonely, and that's a big part of why I open-sourced ReacherX; I wanted people to build with, and the full story is there. When you build alone, there's no one to keep you in check, and the chance of going down a rabbit hole alone is higher.

I've been working on this since mid-2023, and this June I finally went full time on ReacherX.

There's a lot left to do. The idea of telling a tool what people you need and having it connect you with the right people sounds simple. Building it has taken me more than three years, and v4 is still beta. I want people to help improve these foundations and adapt them to their own needs.

Come talk to me

Join the Discord server. Tell me what interests you, or what you tried and couldn't get working. You don't need to arrive with a pull request.

I would like help from people who understand parts of this better than I do. Search, AI agents, background jobs, interface design, testing, documentation: there's work in all of them. There are also decisions I'd like to talk through before writing more code.

Where to look in the code

Start with the architecture overview. It explains the three layers: tools △ Agent can call, workflows that coordinate jobs, and core functions that do the work.

Then pick the part you care about:

Discovery: where possible matches come from.
Qualification: how the system decides whether someone fits.
Enrichment: turning outside data into a useful profile.
Planning and outreach: preparing and carrying out the next steps.
Memory: saved instructions and feedback.
Reporting: where the numbers come from.

Each article points to files you can open. The setup guide covers running a copy locally. If you get stuck, tell me which step failed and what it printed. That helps me fix the instructions too.

Before spending a week on a change

Check the issues and read CONTRIBUTING.md. Please discuss new features, roadmap work, integrations, agent-behavior changes, architecture changes, and major UX changes with me first. A small bug fix or typo doesn't need a long conversation.

For a bug, show how to reproduce it. Explain what changed and how you checked the fix. Screenshots help with interface changes. Search for an existing helper before adding another one.

If coding isn't your thing, try the app with a goal you understand and tell me which results were wrong. Clear feedback, a better instruction, or a walkthrough can help someone else use it.

And if you want to help me keep working on this full time, there's Patreon. You can also star the repository or share it with a friend who might want to build with us.